NHIS Methods

Overview

The National Health Interview Survey (NHIS) is a cross-sectional household interview survey. Cross-sectional surveys collect data asking the same questions to a group of people during the same time period. Each year, NHIS continuously samples and interviews people throughout the year. Sampling is how surveys and other studies choose their participants.

The NHIS sample

The NHIS sample is a probability sample. Probability samples randomly select a small group of participants from a larger group. NHIS selects its sample from all people living across the country—the U.S. population. Potential participants must meet certain criteria to be part of the NHIS sampling.

People sampled to participate in NHIS must live within one of the 50 states or District of Columbia at the time of the interview.

They must live in households or other noninstitutional settings. Noninstitutional settings where people live include homeless shelters, rooming houses, and group homes. People who live temporarily in student dormitories or other temporary housing are sampled within the households where they live permanently.

They must be civilians who don't live in institutions. Civilians are people not serving on active duty in the military. Institutionalized people live in places, like correctional facilities, skilled-nursing homes, and mental (psychiatric) hospitals.

They must have a stable (fixed) household address. Examples of households that don't have fixed addresses include—

• People who are experiencing homelessness and do not live in shelters
• Active-duty military personnel and civilians living on military bases
• People in long-term care institutions like nursing homes
• People in correctional facilities like prisons or jails
• U.S. nationals living in foreign countries

To keep survey operations manageable, cost-effective, and timely, NHIS uses geographically clustered sampling techniques to select the sample of households. This sampling method divides a large, spread-out group (like the U.S. population) into smaller groups or clusters that live in the same defined area (like a county or metro area). Each month's sample is nationally representative to accurately reflect the diversity of the people who live in the United States, and their health, circumstances, and other characteristics.

Redesigning the sampling plan

NHIS redesigns the sampling plan after the census takes place every 10 years. Redesigning the sampling plan ensures NHIS continues to measure and reflect the changing U.S. population accurately. These redesigns also allow NHIS to meet new survey objectives. A new sampling plan for the 2016–2024 NHIS was designed with results of the 2010 decennial Census. A new sample design based on annual sampling and the 2020 decennial Census will be used starting in 2025. Details about this sample redesign will be shared in the future.

The current sample design is not a multistage sample design because the first stage of sampling identifying the geographic clusters, isn't random. The NHIS design uses more than 300 clusters of addresses located in well-defined geographic areas. Geographic areas include a county, a small group of counties that are next to each other, or a metropolitan area. These geographic areas are always located within a single state.

The geographic areas are classified into one to two groups per state. Clusters of addresses that share a defined characteristics (or concept) are then identified using 2010 census housing-unit count information. A sample of address clusters is selected independently for each group for NHIS face-to-face interviews.

The NHIS sample size varies each year. There might be fewer people for budgetary reasons or there might be more people if additional funding is available.

The NHIS questionnaire was redesigned in 2019 to increase data relevance, enhance data quality, and minimize participant (respondent) burden. The wording of some questions and answer choices was revised to improve measurement. Some questions were dropped, and others were added, changing the context for some questions.

The redesign, plus changes in how we developed the new NHIS design and how we weight the data we collect, mean that data collected in 2019 and later cannot be compared to data for earlier years.

Weighting data is a statistical process that adjusts how individual responses and responses from specific groups contribute to the overall picture the data paints. Weighting ensures our statistics and estimates more accurately reflect the entire population.